Multisite musculoskeletal pain in adolescence

(1)

Department of Clinical Medicine, Faculty of Health Sciences

Multisite musculoskeletal pain in adolescence

The relationship with psychosocial problems, mental health and later welfare benefit receipt

—

Christian Eckhoff

A dissertation for the degree of Philosophiae Doctor – May 2016

(2)

(3)

“As you ought not to attempt to cure the eyes without the head, or the head without the body, so neither ought you to attempt to cure the body without the soul. This is the reason why the cure of many diseases is unknown to the physicians of Greece, because they are ignorant of the whole, which ought to be studied also; for the part can never be well unless the whole is

well. It is the great error of our day in the treatment of the human body, that physicians separate the soul from the body.”

The Thracian King Zamolxis.

The Dialog of Charmides, Plato.

(4)

NORSK POPULISTISK SAMMENDRAG (Norwegian summary)

Hvordan går det med smerteplaget ungdom?

En nordnorsk befolkningsundersøkelse.

Fysiske smerter er en utbredt plage blant unge. Ofte finner man ikke en klar fysisk årsak eller sykdom som kan forklare smertene, men vi vet at kroppslige smerter kan henge sammen med stress og psykiske vansker. Formålet med dette forskningsprosjektet var å undersøke

sammenhengen mellom utbredte smerter og psykososiale vansker i ungdomstiden, og undersøke sammenhengen med senere psykiske lidelser og trygdeytelsesbruk fram til 25-års alderen ved kobling til nasjonale helseregistre.

Ungdomshelseundersøkelsen i Nord-Norge er en langsgående undersøkelse som følger ungdom fra hele Nord-Norge fra 10.klasse og 10 år fram i tid. En rekke forhold ved

10.klassingers liv ble undersøkt med spørreskjema i 2003–05. I 2014 koblet vi disse dataene til Norsk pasientregister og Trygderegisteret. Totalt svarte 4881 av 5877 (83 % av

totalpopulasjonen) av ungdommene på spørreundersøkelsen og 3987 av dem (70 % av totalpopulasjonen) godtok kobling til helseregistre.

Utbredte smerter er en vanlig plage hos ungdom, 22,4% rapporterte å ha hatt mye smerter fra tre til fire ulike steder i kroppen det siste året. Jenter rapporterte betydelig mer smerter enn gutter og det var ingen forskjell i smerter mellom samisk og ikke-samisk ungdom. Vi fant en sterk sammenheng mellom utbredte smerter og psykososiale vansker i ungdomstiden hos begge kjønn, der angst/depresjon symptomer, negative livshendelser og skolerelatert stress var de viktigste faktorene.

Vi fant en tydelig sammenheng mellom utbredte smerter i ungdomstiden og en økende forekomsten av senere psykiske lidelser, arbeidsavklaringspenger, sykefravær og sosialstønad hos begge kjønn. Utbredte ungdomssmerter hadde sterkest sammenheng med senere angst og depressive lidelser, hos begge kjønn. Sammenhengen mellom ungdomssmertene og senere psykiske lidelser var i stor grad forklart av deltakernes psykososiale vansker i ungdomstiden.

Utbredte smerter hos unge bør tas på alvor hos begge kjønn. Helsepersonell som møter ungdom med slike plager bør undersøke hvorvidt ungdommen har psykososiale og psykiske vansker som kan tenkes å påvirke smertene. Det kan bidra til å redusere risikoen for vansker senere i livet.

(7)

ACKNOWLEDGEMENTS

The Norwegian Arctic Adolescent Health Study was a joint collaboration funded by UiT The Arctic University of Norway, The Centre for Sami Health Research and the Norwegian Institute of Public Health. The registry linkage with the Norwegian Patient Registry and the National Insurance Agency Registry was funded by the University Hospital of North Norway and Nordland Hospital. I would like to thank all the funders and all the adolescents who participated in this study.

First and foremost, I would like to thank my main supervisor Siv Kvernmo for introducing me to research and for sharing her knowledge with me. Her support, feedback, encouragement and drive has been of great importance to me.

I am also thankful to my co-supervisor Bjørn Straume for statistical advice and for interpretation and planning of paper II and III. I have appreciated our discussions of statistical issues and methods.

A personal interest in this field started when I worked at a psychiatric inpatient ward during my medical studies. I was struck by how the patients lives were influenced by their predicaments, this insight into public health has inspired my thesis. I would like to thank all my previous and current patients for having shown me how painful a troubled life can be.

I would like to thank the Department of Child and Adolescent Psychiatry, Division of Child and Adolescent Health, at the University Hospital of North Norway funding, facilities, flexibility and support and I would like to thank all of my clinical colleagues in the

department for making my writing of the thesis less lonely than it could have been.

A warm thank you to my family and friends for encouragement and patience. A special thank you to Anna; our discussions and mutual support and encouragement has been vital. Thank you also to Anders for fruitful discussions and ideas and to my father for sharing his psychiatric knowledge with me. Last but not least thank you to my mother and sister for always being there, supporting me.

Tromsø, May 2016 Christian Eckhoff

(8)

LIST OF PAPERS

I. ECKHOFF C, KVERNMO S. Musculoskeletal pain in Arctic indigenous and non- indigenous adolescents, prevalence and associations with psychosocial factors: a population-based study. BMC Public Health 2014;14:617.

II. ECKHOFF C, STRAUME B, KVERNMO S. The relationship between multisite musculoskeletal pain in adolescence and mental health disorders in young adulthood – The NAAHS cohort study (Under review).

III. ECKHOFF C, STRAUME B, KVERNMO S. Multisite musculoskeletal pain in adolescence and later receipt of medical and social welfare benefits from adolescence into young adulthood – The NAAHS cohort study (Under review).

(9)

ABBREVIATIONS

AAP Work assessment allowance ANOVA Analysis of variance

ADHD Attention Deficit Hyperactivity Disorder FD-trygd The National Insurance Registry

HSCL-10 Hopkins Symptom Checklist, 10-item version HUNT The Nord-Trøndelag Health Study

IASP The International Association for the Study of Pain ICD-10 International Classification of Disease Version 10

ICPC-2 International Classification of Primary Care, Second edition NAAHS Norwegian Arctic Adolescent Health Study

NPR Norwegian Patient Registry

SDQ Strength and Difficulties Questionnaire

(10)

SUMMARY

Pain is common in otherwise healthy youth and is often associated with psychosocial problems. However, there is limited knowledge of how pain-troubled adolescents fare from adolescence into young adulthood.

To examine this, the Norwegian Arctic Adolescent Health Study, a 10^th grade school- based survey (15–16-year-olds) conducted in North Norway, in 2003–2005, was linked to the Norwegian Patient Registry and the National Insurance Registry. In total, 4,881 out of 5,877 adolescents (83% of the total population) responded to the youth survey, and 3,987 (70% of the total population) consented to the registry linkage. Ten percent of the sample were

indigenous Sami. Musculoskeletal pain was measured by the number of musculoskeletal pain sites (0–4). The aims of the thesis were to explore the relationship between multisite

musculoskeletal pain and psychosocial problems in adolescence, and to determine whether or not adolescent musculoskeletal pain is associated with mental healthcare use and disorders in young adulthood (18–25 years of age). In addition, this thesis examines whether adolescent musculoskeletal pain is associated with later medical and social welfare benefit receipt from adolescence into young adulthood (16–24 years of age).

Similar to previous research, we found multisite musculoskeletal pain to be a common complaint in adolescence, associated with psychosocial and mental health factors in both genders. We found no major ethnic differences; indigenous Sami adolescents were not worse off. The most important adolescent factors associated with musculoskeletal pain were

anxiety/depressive symptoms, negative life events, and school-related stress, which were found in both genders.

We found a significant relationship between the increasing number of adolescent musculoskeletal pain sites and an increasing proportion of later mental healthcare users, mental health disorders, sickness, medical rehabilitation and social welfare benefits, in both genders. Overall, adolescent musculoskeletal pain was not significantly associated with mental health problems in young adulthood when adjusted for adolescent psychosocial and mental health factors. However, adolescent musculoskeletal pain was significantly associated with anxiety disorders, and showing a strong trend toward mood disorders when adjusted for the adolescent factors. Adolescent musculoskeletal pain was a significant predictor of

sickness benefits, in both genders, as well as social welfare benefits, in females, when adjusted for adolescent psychosocial and mental health problems.

Adolescents reporting frequent or recurrent pain may be troubled with psychosocial and mental health problems, and they are at risk of later mental health disorders, as well as health and social difficulties, from adolescence into young adulthood. Healthcare practitioners dealing with adolescents who are troubled by multiple physical symptoms should assess for psychosocial problems and mental health symptoms, offering interventions as necessary.

(11)

1 INTRODUCTION

1.1 About the thesis

This thesis explores the relationship between multisite musculoskeletal pain in adolescents and 1) psychosocial and mental health problems in adolescence, and 2) mental healthcare use and disorders in young adulthood and 3) medical and social welfare benefit receipt from adolescence into young adulthood.

The mind–body causality question has been a classical question throughout the history of philosophy and medicine. It can be traced as far back as to Plato’s dialogs [1]. In the last decades, there has been growing attention to the mind–body causality concept and the underlying mechanisms in modern medicine.

Physical symptoms can be the primary presentation of emotional problems in primary healthcare settings and psychosocial problems are common among youth with unexplained medical symptoms [2]. Up until recent years, epidemiological research on pain focused on single-site pain in relationship with psychosocial problems, while epidemiological studies indicated that multisite pain is common [3, 4], showing a strong linear relationship with different health problems [4]. The lack of epidemiological studies from adolescence into young adulthood was evident concerning the relationship between physical pain and mental health problems [5]. What happens with pain troubled youth and their potential mental health problems in young adulthood? Is adolescent pain indicative of mental health problems over a longer time frame, and would adolescent psychosocial problems influence this relationship?

Adolescence and the early twenties can be a challenging developmental period and is often the time of mental illness debut [6, 7]. Chronic pain and mental health disorders are two major public health issues and research on these issues, especially in youth, can make an important contribution to public health knowledge.

The thesis is based on data from the Norwegian Arctic Adolescents Health Study [8], a cross-sectional population-based study of 15–16-year-olds in the three northernmost counties in Norway (2003–05). In order to examine the relationship between adolescent

musculoskeletal pain and psychosocial and mental health problems into adulthood, the study was linked with the Norwegian Patient Registry [9] and the National Insurance Registry.

1.2 Understanding pain

Pain is a normal and common experience [3]. Pain is the body’s alarm system that indicates that something is wrong. Fundamentally, it is a very useful and sensible system and is

necessary for survival. Unfortunately, pain can become persistent, scary and complex, and the underlying cause can be unclear to the patient, their family and to health practitioners.

Children and adolescents commonly express physical complaints without a clear physical cause [10]. For the most part it is solved by reassurance, but a high proportion of adolescents report chronic pain that disables them in their daily activities [11].

The International Association for the Study of Pain (IASP) defines pain as “an unpleasant sensory and emotional experience associated with actual or potential tissue damage, or described in terms of such damage” [12]. As the definition proclaims, pain is a subjective feeling and no objective test can verify or contradict the degree of the experience.

Acute pain, often due to injury, might be more straight forward with an obvious cause

(12)

compared to chronic pain. Chronic pain, however disabling, might not serve any protective function and can be present even in the absence of a clear physical cause or illness. Pain over time, which can be unclear or maladaptive, is often more associated with the central nervous system [13, 14]. Here, central processes, such as central sensitization, play the most important role in the reduction of inhibitory signals from the brain and amplification of sensory input [15–19]. As Latremoliere and Woolf state, “Pain is not then simply a reflection of peripheral inputs or pathology, but is also a dynamic reflection of central neuronal plasticity” [16].

Pain is a result of complex sensory mechanisms in the nervous system, primarily the brain. Pain and emotion share biological mechanisms

[13, 14, 18, 20–22] and both, as most bodily functions, are a result of biopsychosocial processes [6, 13, 14, 18, 23]. It is well-established that psychosocial factors have an important role in chronic pain [14].

Fig 1 The aspects influencing pain and human processes in general (The biopsychosocial model).

Anxiety or fear, expectations and the social context are important factors in the modulation of pain [13, 14, 18]. All of these factors modulate the pain process (Fig. 2). Ultimately, the brain is responsible for making the final decision of whether the body is in danger and sounding the alarm in the way of pain [13, 14, 18].

Fig 2 An overview of the modern understanding of pain.

Note: Modified version of L. Moseley’s original figure [14].

(13)

1.3 Epidemiological research on adolescent pain 1.3.1 Prevalence

Pain complaints are common in adolescents [3, 24] and tend to persist into adulthood [25–27].

Musculoskeletal pains are most prevalent in adolescents [3, 11, 28–30], showing an

increasing trend in the last decades [31, 32]. Hoftun et al. found that 44.4% of the adolescents in the Young HUNT Study reported one or more pains at least once a week for at least three months, defined as chronic pain [11]. Musculoskeletal pain was most prevalent at 33.4%, followed by headache/migraine. Recurrent [3] and multiple pains [3, 11, 28–30, 33, 34] are also common in adolescence. In the Young HUNT Study 25.5% reported at least two pain sites and 8.5% reported three or more musculoskeletal pain sites [11], supported by high reports of multisite pain in the Norwegian Youth Studies [29] and in studies conducted in other European countries [30, 33, 34].

The self-reported proportions of pain are generally high, but the degree of pain-related functional disability is more unclear and difficult to examine in population studies that

primarily rely on self-reports. Hoftun et al. found that 80% of the adolescents reporting chronic pain also described having one or more pain-related disabilities and there was an increased proportion of disability in those reporting multisite pain [11]. In adolescents reporting increasing mental health problems or disorders, chronic multisite pain with high disability is very prevalent [35, 36].

Fig 3 The proportion of multisite pain (three or more pain sites) in adolescents (%).

Note: Young-HUNT study: Pain at least once a week during the last 3 months [37]. Northern Finland Birth Cohort: Pain during the last 6 months [33]. Youth Studies: Often troubled by pain during the last 12 months [29].

1.3.2 Sociodemographic factors

Pain reports are generally higher in females than in males and increases with age [3, 11]. In the Young HUNT Study, 40.0% of females reported chronic musculoskeletal pain compared to 26.5% of the males [11]. The proportion of multiple chronic musculoskeletal pain was

0 5 10 15 20 25 30 35 40 45 50

3 months (Young-HUNT) 6 months (Northern

Finland Birth Cohort) 12 months (Youth Studies)

The proportion of adolescents (%)

Females Males

(14)

10.9% in females and 6.1% in males, respectively [11]. Paananen et al. found that 39% of the females reported three to four musculoskeletal pain sites during a six month period compared to 23% of the males [33].

Research on the relationship between parental socioeconomic status and pain in youth has shown conflicting findings, but overall, King et al. concluded that low socioeconomic status was associated with increased pain complaints [3]. In some studies, lower parental education was associated with higher pain reports [38, 39], but other studies have found no such association [40, 41]. Low family income [38, 42] and parental unemployment are shown to be associated with increased pain reports [42].

Few studies have examined the ethno-cultural differences in physical pain and potential psychosocial differences [10, 24]. There have been no prior pain studies among the Sami, the indigenous population in Scandinavia, or in other indigenous adolescent groups.

1.3.3 Physical factors

Pain can be a result of muscular aches due to both vigorous activity and inactivity, and prior studies have shown conflicting results. There are reports of higher pain prevalence in

adolescent groups reporting vigorous exercise [30, 33, 43–45], while others found no

relationship with the level of physical activity [37, 46, 47]. An association between sedentary activity and pain reports has been found [33, 37, 44, 45], primarily in females [37, 44]. In addition, there is an indication of increased pain reports in adolescents with a higher BMI [30, 33, 40, 45].

1.3.4 Psychosocial factors

Adolescence may be a stressful period and several psychosocial difficulties are associated with adolescent physical complaints and pains. The high prevalence of physical complaints is considered to be a possible manifestation of psychosocial problems [2], or at least is

potentially influenced by them. The minority of patients referred to pediatric outpatient clinics with medically unexplained chronic pain have somatic disorders that explain the pain [48].

Social support and positive relationships are important supportive factors. Physical complaints have been found to be associated with parental and peer problems [10, 49–53], poor resilience [35, 54] and low self-esteem [35, 50]. Parental support is important for the social functioning of adolescents with chronic pain [55], and parental chronic pain is

associated with pain [56], psychosocial and mental health problems in their offspring [57, 58].

Stress, pressure and negative life events, known risk factors of mental health problems [6], are also associated with pain and other physical complaints [10, 46, 52, 53, 59–63].

Multiple somatic complaints impact daily life and function, and are associated with school difficulties and absence [49, 59].

In a clinical study, a model that included social difficulties, family structure, parental somatization and school problems predicted psychiatric morbidity in pediatric patients with medically unexplained chronic pain [48]. All the above mentioned findings emphasize the need for a dynamic understanding of pain in youth. Physical problems, such as chronic and/or recurrent pain, can have an impact on all social aspects of life and vice versa.

(15)

1.3.5 Mental health

The debut of most mental health disorders is during the 12–24 year age range [6].

Adolescence and young adulthood is the time when most people establish social relationships, academic careers and enter the job market. Mental health disorders can potentially have a negative impact on all of these areas [6]. Previous studies in youth have shown a strong relationship between pain, and other physical complaints, and mental health problems, and this relationship was found in both cross-sectional [30, 35–37, 41, 45, 46, 48, 59, 64–70] and in longitudinal studies [29, 33, 34, 43, 47, 71–79]. These findings are comparable to research in adults [80–84]. Overall, the longitudinal studies have shown a bidirectional relationship between pain and mental health problems in youth [29, 33, 34, 43, 47, 71–79]. Except for the recent findings Shanahan et al. [77] presented, most of the studies use a shorter timeframe [5]

and do not examine this relationship beyond adolescence.

Up until recent years, pain studies focused on single-site pain in relationship with psychosocial and mental health problems. More recent research has shown that the

association between pain and mental health problems is stronger for multisite pain compared to single-site pain [4, 30, 34, 35, 45, 75, 85]. As poor mental health is associated with several health concerns [6], an increased focus on multisite pain in relation to psychosocial and mental health problems is warranted [5].

Having multiple physical symptoms is strongly associated with mental health disorders, especially mood and anxiety disorders, influencing the clinical picture of these disorders [20, 36, 68, 82, 86–89]. Comorbid physical complaints are a common way of presenting mental health problems [10, 65, 86, 88, 89] and are important signs in the early detection of mental health problems [2, 68, 86]. Anxiety and mood disorders have been found in about 27% of patients with chronic pain complaints in primary healthcare [86]. Mental health problems are prevalent in patients referred to pediatric outpatient clinics with medically unexplained chronic pain [48, 74]. The majority of the patients continue to have pain and many struggle with mental health problems years later [74]. A matter of concern is the low recognition of mental health disorders, which are associated with high levels of physical symptoms, among adolescents in primary care [68].

Shanahan et. al.’s recent findings show that pain in children and adolescents is predictive of mood and anxiety disorders in young adulthood, and youth with persistent pain are at increased risk of later mental health problems [77]. This suggests that the relationship between pain and mental health problems in youth extends beyond adolescence. However, the influence of potential adolescent mediators needs to be examined further [77].

1.3.6 Medical and non-medical welfare benefit receipt

Mood and anxiety disorders are the most prevalent mental health disorders resulting in long- term medical benefits in Norway [90], and the mental health disorders with the strongest relationship with adolescent musculoskeletal pain [36]. Mental health problems in

adolescence and young adulthood have been found to be associated with reduced workforce participation and increased medical welfare benefits in young adulthood [91–95]. Adolescent health complaints and single-site pains have been shown to be associated with long-term medical benefits in young adulthood [91, 96]. Studies in adults have shown that multisite pain predicts long-term medical benefits, both overall [97] and those due to mental health disorders

(16)

[98]. A potential relationship between adolescent pain and the later receipt of medical and non-medical benefits has only been sparsely examined.

Young adulthood is an important period where major life struggles can lead to long- term benefit dependency [99, 100]. Continuous sickness benefits of more than eight weeks have been found to be predictive of later disability pension, especially due to mental health and musculoskeletal disorders, in Norwegians under 40 years of age [101]. Over the last decades, there has been an increase in long-term medical benefit recipients due to mental health disorders in 18–29-year-olds in Norway [90, 102]. Mental health disorders are the reason for medical rehabilitation benefits in two out of three cases in those under the age of 30, with a 30% increase from 1996–2008 primarily due to mood and anxiety disorders [90].

Chronic pain and mental health disorders, two major public health issues, are costly to the quality of life for young people [11, 103–105], to their families [103, 106] and to society [6, 97, 105, 107–109]. It is known that pain trajectories start early in life and the estimated economic healthcare costs of chronic pain in youth and adults are substantial [5, 109, 110].

Therefore, early detection and interventions are of major importance. Research on these issues, especially in youth, can make an important contribution to public health work.

(17)

2 AIMS OF THE THESIS

The overall aim was to explore the relationship between multisite musculoskeletal pain in adolescence and adolescent psychosocial and mental health problems, mental healthcare use and disorders in young adulthood and medical and social welfare benefit receipt from adolescence into young adulthood. To explore this, we linked the Norwegian Arctic

Adolescents Health Study [8], a population-based study, with the Norwegian Patient Registry [9] section on specialized mental healthcare and with the National Insurance Registry.

Fig 4 Study design and hypothetical study relationships.

The specific aims of the research papers were as follows:

Paper I – Prevalence of multisite musculoskeletal pain in adolescence and the association with adolescent psychosocial and mental health problems

1 To examine the prevalence of multisite musculoskeletal pain in indigenous and non- indigenous Arctic adolescents.

2 To examine the association with physical, psychosocial and mental health problems in a hierarchical model and determine the importance of the factors.

3 To examine the importance of pain-related functional impairment on this association.

Paper II –The association between adolescent musculoskeletal pain and mental healthcare use and disorders in young adulthood

1 To examine if multisite musculoskeletal pain in adolescence was associated with mental healthcare use and disorders in young adulthood.

2 To determine the importance of adolescent musculoskeletal pain in relation to later mental healthcare use and mental health disorders, when adjusting for sociodemographic factors, as well as adolescent psychosocial and mental health problems.

3 To explore any potential difference in adolescent predictors of later mood, anxiety and developmental and behavioral disorders.

Sociodemographic factors

Musculoskeletal pain and physical factors

Psychosocial factors

Mental health factors

Medical and social welfare benefits

Mental health problems

Adolescent factors Adolescence and young adulthood

(18)

Paper III – The association between adolescent musculoskeletal pain and medical and social welfare benefits from adolescence into young adulthood

1 To examine if multisite musculoskeletal pain in adolescents was associated with an increase in medical and social welfare benefits receipt from adolescence into young adulthood.

2 To determine if adolescent musculoskeletal pain uniquely contributed to the prediction of later sickness and social welfare benefits when adjusting for sociodemographic,

adolescent psychosocial and mental health problems.

3 To explore gender differences in adolescent predictors of later sickness and social welfare benefits.

(19)

3 MATERIAL AND METHODS

3.1 Study design and sample

The Norwegian Arctic Adolescent Health Study (NAAHS) [8] was linked to the Norwegian Patient Registry (NPR) and the National Insurance Registry in 2014. Paper I describes cross- sectional findings from the NAAHS, Paper II and III describes longitudinal findings from the registry linkage.

Fig 4 Study timeline.

Fig 5 Study design and sample (*of the total population, **of NAAHS participants).

The Norwegian Arctic Adolescent Health Study

The NAAHS was conducted among 10^th graders (15–16-year-olds) in nearly all junior high schools (292 out of 293) in the three northernmost counties in Norway, in 2003–05. The questionnaires were administered in classroom settings, monitored by project staff, and

2003 2004 2005 2006 2007 2008 2009 2010 2011 2012

National Insurance Registry (2003–11)

Norwegian Patient Registry (2008–12) NAAHS (2003–05)

15–16-years-old 23–25-years-old

NAAHS- Registry linkage

n=3987 (70%*) NAAHS 10^th grade Northern Norway

(2003–05) n=4881 (83%*)

National Insurance Registry (2003–11) Norwegian Patient

Registry (2008–12)

Registry linkage refusal n=894 (18%**)

(20)

completed during two school hours. Few students used more than 45–60 minutes. Students who were not present in class during the questionnaire administration completed the questionnaire at a later date. The students and their parents were given written information about the study, and the students provided written consent. There were no specific exclusion criteria in this study.

The Norwegian Patient Registry

The NPR [9], a detailed registry from 2008, includes personal identification of specialized healthcare utilization and diagnosis. We used available data from specialized mental

healthcare from 2008 through 2012 when the participants were 18–20 to 23–25 years of age.

The National Insurance Registry

The National Insurance Registry (FD-trygd) keeps records of national medical and non- medical welfare benefits, and we used available data from 2003 through 2011. Follow-up time started at July 1^st of the corresponding year the participants responded to the NAAHS, and had finished Norwegian junior high school. Resulting in a total follow-up time of 6.5–8.5 years (from 16 to 22–24 years of age).

In Norway, citizens can receive medical benefits for impairing illness resulting in reduced working ability. Medical benefits include sickness benefits (up to 52 weeks for employed citizens), medical and vocational rehabilitation, and temporary and permanent disability pension. Medical and vocational rehabilitation and temporary disability pension were organized into one single benefit in 2010, called AAP (work assessment allowance).

Citizens have to be at least 18 years old to be eligible for AAP and disability pension, which requires 50% reduced work ability due to disease, injury or impairment. Social welfare benefits can be received by people who do not qualify for any other benefits and who are unable to financially support themselves.

Sample

In total, 4,881 of the 5,877 (83% of the total population) invited 10^th grade students responded to the NAAHS, of whom 50.1% were females, and 10% were indigenous Sami. In the non- Sami group 64 adolescents reported having other nationalities, thus consisting mainly of majority Norwegians. The following response rates were 88% for Nordland, 82% for Troms and 71% for Finnmark. The lower response rates in Troms and Finnmark may be due to the fact that seven out of 74 schools in Troms and 12 out of 54 schools in Finnmark were not able to facilitate the health study at school, and the questionnaires were therefore mailed home to the students. All these schools had a low number of students. In Nordland one school refused to participate.

Of the NAAHS participants 3,987 (82%) consented to a future registry linkage, resulting in a 70% sample of all 15–16-year-olds in Northern Norway. The registry sample consisted of 49.9% females and 9.2% indigenous Sami.

In order to examine the representativeness of the proportion of mental healthcare users in our sample (30% nonresponders), the NPR calculated the total number of mental healthcare

(21)

Norway public database (n=5,715) to give and approximate cumulative prevalence of 14.9%

mental health care users in the total population, which we compared our sample to.

3.2 Adolescent study measures

3.2.1 The Norwegian Arctic Adolescent Health Study and Statistics Norway Adolescent musculoskeletal pain and physical factors

Musculoskeletal pain was measured by “yes/no” answers to the question: “During the last 12 months have you often been troubled by pain in the head, neck/shoulder, arms/legs/knees, abdomen or back?” Abdominal pain was excluded due to the potential confusion with menstrual pain. Headache was included based on its frequent co-existence with

musculoskeletal pain and their shared mechanisms and risk factors [111]. Resulting in a variable ranging from zero to four pain sites.

Pain-related functional impairment was present if the participants reported reduced activity during leisure time due to pain (yes/no). Those reporting functional impairment due just to abdominal pain were excluded (n=23 in the registry sample).

Physical activity was measured by the question: “How many hours per week do you spend on physical activity, to an extent that makes you sweat and/or out of breath?” Possible answers: 0, 1–4, 5–7 and ≥8 hours per week [112].

Self-rated health was measured by the question: “How is your health right now?”

with four possible options dichotomized into “not good/not so good” and “good/very good”.

Physical injury during the last year was measured. Originally, the participants were asked if they had experienced a serious illness or injury during the last year, and were asked to describe their illness/injury. There were few reports of illness/injury of serious nature and the number of chronic illnesses reported was too low for statistical analysis. The number of physical injuries, mostly extremity injuries and some concussions, were 42.6% (n= 136) of the total responses.

Sociodemographic factors

Socioeconomic status: Information about the participants parent’s occupation was obtained and classified according to the International Standard Classification of Occupation, ISCO-88 [113], and later reclassified into five categories based on the parent with the highest rated occupation. Parental work reported “unknown” was recoded into the missing group.

Parental education: Parents’ highest education was obtained from Statistics Norway’s education registry, registered when the participants were 15–16 years old. Parental education was categorized from “lower secondary” (≤10th grade), “upper secondary” (≤13th grade),

“Lower university degree” (up to 5 years) to “higher university degree” (more than 4 years) [114].

Family income: Adolescents reported their family’s economic situation compared to other families on a four-point scale from “not well off” to “very well off.”

Sami ethnicity was measured by participants having one or more of the following factors: Sami parentage, Sami language competence in parents, grandparents and the participants, and Sami ethnic self-labeling [115].

Adolescent psychosocial factors

(22)

Resilience was measured by a five-item version (α=0.77) of the General perceived self- efficacy scale [116] with higher scores indicating higher resilience. In the five-item version questions one, two, four, seven and nine were used, and scored on a four-point Likert scale from “completely wrong” to “completely right.”

Parental involvement was measured by a four-item version of the Parental

Involvement Scale (α=0.78) [117]. Based on the questions: “My parents know where I am at and what I do in the weekend,” “my parents know where I am and what I do on weekdays,”

“my parents know who I spend my leisure time with” and “my parents like the friends I spend time with.”

Parental support (α=0.88) was measured by the following five statements: “I feel attached to my family,” ”my family takes me seriously,” “my family values my opinions,” ”I mean a lot to my family” and “I can count on my family when I need help.”.

Peer support (α=0.84) was measured by the following four statements: “I feel closely attached to my friends,” “my friends value my opinions,” “I can help/support my friends,”

and “I can count on my friends when I need help”.

Parental involvement, parental and peer support were scored on a four-point Likert scale from “completely agree” to “completely disagree.” Higher scores indicating more problems.

School-related stress (α=0.66) was measured by the following experiences: “Have you ever experienced any of the following:” “Heavy work pressure at school,” “heavy pressure from others to succeed/do well at school,” “find it very difficult to concentrate in class” and

“find it very difficult to understand the teacher when he/she is teaching?” Responses were scored on a three-point Likert scale from “no” to “yes, often.”

Negative life events (α=0.55) were measured by 12 dichotomized questions on different categories of difficult life experiences: “Have you in the last 12 months had anyone of the following problems,” “conflict or fights with your parents,” “parental mental health problems,” “parental financial problems,” “parental drug problems” or “peer problems?”

Responses were measured on a four-point Likert scale from “no, never” (0), “yes, sometimes”

(1), “several times” (2), to “very often” (3). Furthermore, respondents were asked, “have you in the last 12 months experienced trouble being bullied at school/ on the way to school?” with the following options: “never” (0), “sometimes” (1), “about once a week” (2), and “several times a week” (3). Also, “Have you in the last 12 months been exposed to violence?” with the following options of “never” (0), “yes, only by adolescents” (1), “yes, only by adults” (2), and

“yes, by both adolescents and adults” (3). Lastly, respondents were asked, “have you in the last 12 months experienced the following:” “parental unemployment or social care,” “serious illness or injury to yourself,” “serious disease or injury to someone close to you,” “death to someone close to you” or “sexual assault?”

For peer problems (α=0.52) we used the Strengths and Difficulties Questionnaire (SDQ) [118], that measures problems during the last six months.

Adolescent mental health factors

Mental health was examined by anxiety/depression symptoms measured by the Hopkins

(23)

24, and for Sami and non-Sami [120, 121], with a cut-off of 1.85 indicating a presence of emotional distress.

For hyperactivity (α=0.64) and conduct problems (α=0.47) we used the Strengths and Difficulties Questionnaire (SDQ) subscales [118], that measures problems during the last six months. For paper III these two scales were summed into externalizing problems (α=0.69) [122].

Help seeking was measured by mental health help seeking behavior during the

previous year (yes/no) and the use of a psychiatrist/psychologist during the previous year was dichotomized from the responses “no,” “1–3 times”, and “4 or more times”.

3.3 Registry outcome measures

3.3.1 Mental health in young adulthood: The Norwegian Patient Registry

Mental healthcare users consisted of participants found in the specialized psychiatric patient registry, including use of public psychiatric healthcare and private specialists. We

constructed an ordinal variable of “not a patient,” “outpatient only” and “inpatient.”

The number of outpatient contacts and inpatient admissions, and the sum of all outpatient treatment hours and inpatient treatment days was calculated.

Mental health disorders: Each participant’s primary and secondary diagnoses were organized according to the main chapters in the ICD-10 [123]. We used a classical model for psychiatric diagnoses to achieve theoretically constructed groups of reasonable size. We recorded whether the participants had received a diagnosis from any of the five diagnostic groups: substance use disorders (F10–19), psychotic (F20–29), mood (F30–39), anxiety (F40–

49), developmental and behavioral disorders (F50–98), and undiagnosed. We included both primary and secondary diagnoses due to an evident difference in diagnostic coding practice, making it difficult to pick out the primary disorder in patients with several diagnoses.

3.3.2 Welfare benefits from adolescence into young adulthood: The National Insurance Registry

Medical insurance benefits

Sickness benefits were measured as a continuous variable on total 100% sickness days during the registry period. Graded sick leave days were recalculated into 100% sickness days.

Sickness benefits were also organized into the proportion of participants with <1, 1–3, 3–6 and 6–12 months of benefits in a 12-month period, not within a calendar year.

Sickness benefits due to mental health included disorders primarily from the ICPC-2 [124]. Primarily three diagnostic codes were used: P76 (Depressive disorder, 42.1%), P02 (Acute stress reaction, 26.2%), and P29 (Psychological symptom/complaint other, 12.0%).

Sickness benefits due to musculoskeletal problems from the ICPC-2 were dichotomized. Musculoskeletal problems due to injuries (L72-81) were excluded.

Medical rehabilitation benefits, overall and due to mental health problems, and disability pension were recorded.

Non-medical benefits

Social welfare benefits were measured as a continuous variable by the total number of months with social welfare benefits. The mean of total social welfare months was converted from

(24)

months into days in Figure 1 and Table 1 and 2 (Paper III) for comparison purposes with sickness days. Due to being recorded at a monthly basis social welfare benefits were organized into 1–3, 3–6 and 6–12 months of benefits in a 12-month period.

3.4 Statistical methods

The general thought for the methodology of this thesis was to examine the associations with the outcomes in more whole models controlling for relevant factors, while trying to minimize over-modeling. In general, the adjusting factors were controlled for in a hierarchical fashion following their respective characterized group.

Fig 6 The methodological groups of study factors.

We examined the adolescent factors to see whether the participants refusing the registry linkage differed from those who accepted. We found no significant difference between the registry sample and the missing sample except for slightly lower mean of negative life events in the missing sample. Overall, the missing sample was not worse off.

Chi-square tests and one-way ANOVA were used for univariate analyses. Anxiety/depression (HSCL-10) was handled continuously in all multivariable analyses. All analyses were

conducted with SPSS 21–23 (IBM software).

Paper I – Prevalence of multisite musculoskeletal pain in adolescence and the association with adolescent psychosocial and mental health problems:

Hierarchical, backward, linear multiple regression analysis was used for analysis of

significant univariate predictors. Stratified analyses for gender, ethnic (Sami vs. non-Sami), and pain-related functional impairment were carried out. First, sociodemographic factors were analyzed, followed by physical, psychosocial, and mental health problems. The statistical significance level was set to .01 due to the large number of participants, except for the Sami group where .05 was chosen due to lower N.

Paper II – The association between adolescent musculoskeletal pain and mental healthcare use and disorders in young adulthood:

Initially we examined the prediction of later mental healthcare use by ordinal and multinomial regression. However, we found no linearity between the constructed ordinal groups (not a patient – outpatient only – inpatient). The main statistical difference was between the users and nonusers of mental healthcare, and not between the user groups (outpatient only vs.

Musculo- skeletal pain and physical

factors

Sociodemo- graphic

factors

Psychosocial

factors Mental health factors

(25)

Hierarchical logistic regression was used for the multivariable analysis for later mental health care use (Table 3). In Step 1 the sociodemographic (Model 1), physical factors (Model 2: musculoskeletal pain adjusted for sedentary and physical activity) and psychosocial factors (Model 3, 4 and 5) were analyzed grouped together in models based on their respective characteristic groups. Insignificant factors were not included in the next steps in order to simplify the models. In Step 2, the significant sociodemographic factors from Model 1 were added as adjustments for musculoskeletal pain alongside physical activity (Model 6). The significant psychosocial factors from Step 1 were analyzed together in Model 7. In the final model, the significant psychosocial factors from the second step (Model 7) were added to the adjustment of adolescent musculoskeletal pain.

Hierarchical logistic regression was used for the multivariable analyses on the diagnostic groups of mental health disorders (Table 4), following the same model building approach as described above for Table 3. However, only the significant factors from Step 1 in the mental healthcare use model were examined in order to simplify the models. In addition, musculoskeletal pain, physical activity and the sociodemographic factors were included in the final model regardless of if they were found insignificant in Step 1 and 2 (Model 1, 2 and 6), while insignificant psychosocial factors from Model 7 were excluded.

The statistical significance level was set at .05. Bonferroni-adjusted significance levels for multiple comparisons (.05/nfactors) are presented for the “final model” in the multivariable analysis tables (Tables 3 and 4) as a conservative comparison of significance.

Paper III – The association between adolescent musculoskeletal pain and medical and social welfare benefit receipt from adolescence into young adulthood:

Linear regression was used in the multivariable analyses on the total length of sickness and social welfare benefits (Tables 3 and 4). First, multivariable linear regression analyses with the sociodemographic, physical, psychosocial and mental health factors were analyzed in their characterized groups. The insignificant factors were excluded and the significant factors were used to adjust for adolescent musculoskeletal pain (Tables 3 and 4). Gender and parental education level were included in the models regardless due to adjustment reasons. In Tables 3 and 4 “Model 1” presents the adjustment for sociodemographic factors, “Model 2” with the addition of adolescent psychosocial factors, “Model 3” for adolescent mental health problems.

To explore for gender differences “Model 3” was stratified by gender. Adjustments for follow-up time and counties did not alter the multivariable findings and are not presented in this paper.

The statistical significance level was set to .05. Bonferroni-adjusted significance levels for multiple comparisons (.05/nfactors) are presented for “Model 3” in the multivariable

analysis tables as a conservative comparison of significance.

(26)

4 ETHICS

The Norwegian Data Inspectorate and the school authorities approved the NAAHS. The Regional Medical Ethical Committee approved the NAAHS and the registry linkage. The Norwegian Institute of Public Health and Statistics Norway carried out the registry linkage.

The students and their parents were given written information about the NAAHS, and the students provided written consent. The participants accepted future registry linkage at the time of the NAAHS. The participants had the opportunity to remove their consent at a later date.

Data from the NPR is used in this thesis. The interpretation and reporting of these data are the sole responsibility of the authors, and no endorsement by the NPR is intended nor should it be inferred.

(27)

5 RESULTS – SUMMARY OF RESEARCH PAPERS

5.1 Paper I – Prevalence of multisite musculoskeletal pain in adolescence and the association with adolescent psychosocial and mental health problems

The prevalence of musculoskeletal pain and mental health problems was significantly higher in females, except for arm/knee/leg pain, as was the number of musculoskeletal pain sites. In total, 29.1% of the female and 15.8% of the male adolescents reported three to four

musculoskeletal pain sites. A total of 38.3% of the adolescents reported pain-related

functional impairment. We found a strong association between musculoskeletal pain sites and adolescent psychosocial and mental health problems. Of the adolescents reporting zero or one musculoskeletal pain sites, 10% scored above the anxiety/depression measure cut-off (HSCL- 10 >1.85), indicating emotional distress, compared to 36.5% for the adolescents reporting four pain sites. Anxiety/depression, negative life events, and school-related stress were the most important factors associated with musculoskeletal pain, especially in those reporting pain-related functional impairment. The gender difference in pain reports was due to

increased reports of psychosocial problems in females. We found no major ethnic differences.

5.2 Paper II – The association between adolescent musculoskeletal pain and mental healthcare use and disorders in young adulthood

The proportion of mental healthcare users was 13.6% (59.5% were females). Multisite adolescent musculoskeletal pain was significantly associated with an increase in mental healthcare use and mental health disorders in young adulthood. The relationship was stronger for anxiety and mood disorders in both genders. Of the participants reporting zero or one musculoskeletal pain sites in adolescence, 10.3% were registered as mental healthcare users in young adulthood compared to 20.2% in those reporting three to four pain sites in adolescence.

A higher proportion of mental healthcare users was found in the participants reporting pain- related functional impairment – 17.6% compared to 12.2% in the non-impaired. We found a three times increase in outpatient hours and a 3.7 times increase in the number of mental healthcare inpatients from zero to four adolescent pain sites. Overall, the association between adolescent musculoskeletal pain and later mental health problems was mediated by adolescent psychosocial and mental health problems, not by physical or sedentary activity. However, when we examined the different mental health disorders, we found musculoskeletal pain to be significantly associated with anxiety disorders, with a strong trend towards mood disorders, after adjusting for the adolescent factors.

5.3 Paper III – The association between adolescent musculoskeletal pain and medical and social welfare benefits from adolescence into young adulthood The proportions of sickness benefit recipients, total number of sickness benefit days and sickness benefits due to mental health problems were higher in females. There was no gender difference for social welfare benefits. We found a linear relationship between adolescent musculoskeletal pain sites and the occurrence of sickness, medical rehabilitation and social welfare benefits in young adulthood. The participants reporting zero musculoskeletal pain sites in adolescence had received 17.1 sick days and 33.6 days of social welfare benefits on average, compared to 52.4 sick days and 93.9 days of social welfare benefits in those

(28)

reporting four pain sites in adolescence. The total amount of sickness and social welfare benefits was significantly higher in those reporting pain-related functional impairment

compared to the non-impaired. Adolescent musculoskeletal pain was a significant predictor of sickness, in both genders, and social welfare benefits in females when adjusted for adolescent psychosocial and mental health problems. The most important adolescent psychosocial predictors were externalizing problems, less parental involvement and negative life events.

(29)

6 DISCUSSION

6.1 Methodological issues 6.1.1 Study design and sample

The main strength of this study is the linkage of a large population-based study to a patient registry. It is the first and only study inviting all 10^th graders in Northern Norway to participate, including indigenous Sami adolescents. Only one high school refused to

participate in the study. We had a high participation rate in the cross-sectional study and the registry linkage, and an equal gender distribution. Both aspects strengthened the validity and generalizability of the study. The 13% that refused the registry linkage did so at the age of 15–16 when they completed the questionnaire. This reduced the chance of bias in the registry outcomes as those that are seriously ill are more prone to be lost during the follow-up.

Approval at a later date might have reduced the participation rate; however, all participants were informed that they could withdraw from the study at a later date if desirable. No participants have done this to date.

The registry sample was representative of the total sample and the participants refusing the registry linkage were not worse off. We had a representative sample of mental healthcare users in our sample, which was only 1.3% lower compared to the total population estimation (13.6% compared to 14.9%). A calculated estimation of mental healthcare users in our missing sample was 17.8%, supporting the notion that non-responders are more troubled.

The higher proportion of mental healthcare users in the missing sample supports a mild selection bias, which is difficult to avoid in studies requiring consent. We compared the number of sickness benefit recipients in our sample to data from the statistical bank of Statistics Norway [125]. Sickness benefit recipients from the three northernmost counties in the first quarter of 2011, according to Statistics Norway, showed a 5.2% sickness absence compared to our 4.0%, indicating increased sickness benefits in the non-responders.

In light of these findings, the results of this study are representative of Norwegian adolescents. The relationship between adolescent musculoskeletal pain and mental health problems, and later medical and social welfare benefit receipt, is probably generalizable to other modern societies. Research from different nations has found a relationship between adolescent pain and later mental health problems, supporting this claim [3, 30, 33, 37]. Still, caution should be used as societal, cultural, socioeconomic and health service differences could influence these findings.

We only had one cross-sectional study linked to the registries, and regarding the timeframe of the study, there might be other factors influencing the associations found in this study. The lower explained variance of the regression models in Papers II and III can lead us to question the true impact of the significant adolescent factors. However, explained variance is a relative value, dependent on the nature of the true relationship examined. Repeated measures in adolescence may have singled out more troubled individuals compared to our single population study.

6.1.2 Statistical methods and precision

Epidemiological research depends on statistical regression models to explore associations in a population sample that hopefully resemble a true relationship found in the true population.

(30)

Psychosomatic problems are complex and the relationship between physical and mental health is truly a two-way street. The potentially influencing factors are many, while there is an upper limit to the complexity of regression models that can be derived with any acceptable degree of uncertainty [126]. Naturally, there are more important factors that are possible to highlight in a population sub-sample.

One goal of this thesis was to adjust for several relevant mediators found to be associated with pain and mental health problems. We had a large sample, with sufficient power to perform multiple comparisons. A hierarchical model was chosen as we wanted to show the dynamics of how each characteristic group of adjusting factors (sociodemographic, physical, psychosocial and mental health factors) influenced the relationship between

adolescent musculoskeletal pain and the respective outcomes. Insignificant factors were generally excluded in the hierarchical steps up to the final models in order to reduce the complexity of the models and highlight the more important factors.

Bonferroni corrections were carried out to adjust for multiple comparisons in the multivariable regression models of Papers II and III. The Bonferroni correction is a

conservative adjustment, adjusting the p-value at which the test is evaluated for significance based on the total number of tests being performed [127]. It targets Type 1 errors, reducing the number of incorrect rejections of the null hypotheses, at the expense of Type 2 errors and the power to detect a relationship. The Bonferroni adjustment was included to distinguish the more reliable associations from the more questionable. In Paper I, a significance level of 1%

was chosen to reduce Type 1 errors.

6.1.3 The Norwegian Arctic Adolescent Health Study

The population study relied on self-reports with the risk of information bias. The specification of a 12-month period for the pain questions may have increased the risk of recall bias

compared to shorter time periods. The expression “several times” is objectively vague and is open for interpretation, but it indicates some regularity and seriousness of pain. The yes/no answers did not offer any further possibilities to differentiate the seriousness of the pain and a more stringent measure of pain, such as in the Young HUNT studies [11], might have resulted in stronger associations. However, by focusing on the number of adolescent pain sites the importance of multisite pain was still emphasized. Naturally, somatic illness could be the cause of multisite pain, but excluding a somatic cause for the pain was not possible due to the nature of the survey. In an attempt to adjust for somatic illness, we explored the descriptions of serious illness or injury during the previous year (n=366). We found a few reports of serious or chronic illness (for example, Juvenile Rheumatoid Arthritis and migraine), probably caused by underreporting, perhaps due to the subjective interpretation of the term

“serious”. Most responses were normal physical injuries (extremity fractures and sprains).

The psychosocial and mental health factors in the population study included more commonly used and validated scales, such as the HSCL-10 [119, 120] and the SDQ-scales [118], and measures less frequently used outside of the Youth Studies [8]. The use of

validated instruments reduces the chance of measurement error by increasing the validity and objectivity of the measure [128]. However, the HSCL-10 only measured anxiety/depression

(31)

timeframe for the psychosocial measures, with the exception of the SDQ-scales and the negative life events measure. Still, most scales had a high internal consistency.

6.1.4 The Norwegian Patient Registry

Fig 7 Mental healthcare users and disorders (percentage of the total population)

The NPR is a high quality national patient registry for specialist care; we found few errors.

Ideally, it would have been preferable to differentiate between primary and secondary

diagnoses, but from a close inspection of the NPR data, it was clear that attempting this would not be trustworthy due to the evident difference in the specialists’ diagnostic coding practice.

However, most of the primary and secondary diagnoses were within their respective main chapters of the ICD-10. A natural skepticism must be kept due to the specific diagnostic coding by health practitioners. Diagnostic coding can be a challenge, both in psychiatric and somatic care, and the practitioners’ coding tradition can influence the choice of a specific diagnosis. As a result, there is a risk of misclassification in terms of specific diagnoses, a risk in all diagnostic psychiatric research. However, compared to a clinical study, we had no way to ensure that the clinicians were “on the same page”.

6.1.5 The National Insurance Registry

Fig 8 Medical and social welfare benefits (percentage of the total population)

Mental healthcare users 2008–2012 N=543 (13.6%)

Substance use (F10–19) N=64 (1.6%) Psychotic (F20–29)

N=28 (0.7%) Mood (F30–39)

N=165 (4.1%) Anxiety (F40–49)

N=201 (5.0%) Developmental and behavioural (F50–98)

N=130 (3.3%) Undiagnosed N=129 (3.2%)

(32)

The National Insurance Registry had more errors compared to the NPR, primarily in the diagnostic coding, where 61% of disability pensions, 12% of medical rehabilitation and 21%

of the sickness benefits recipients had no registered diagnosis. The diagnostic challenges were even greater for the medical insurance benefits compared to the NPR. Three diagnostic codes were primarily used by the general practitioners for sickness benefits due to mental health problems, and a few were due to anxiety disorders (5.5%), the most prevalent mental health disorder. Consequently, we did not differentiate between the ICPC-2 diagnoses for mental health problems. Most likely, the diagnoses for medical rehabilitation and disability pension were less prone to misdiagnosis due to the seriousness and the time aspect of the illness.

Contrary to previous studies where six months of sickness benefits were dichotomized with medical rehabilitation and disability pensions [91, 92, 96], we explored the outcomes on multiple levels. Since only shorter time periods, <1–3 months, of benefit reception was not significantly associated with adolescent musculoskeletal pain, we found it more reasonable to use the total amount of benefits received over the entire study period. We also recalculated graded sickness days into 100% sickness days, contrary to just counting 100% sickness days [96], increasing the validity of the measure. Some sick leave periods were missing a date for the completion of the benefit period. We used the last registered date for the benefit, a sick leave extension date, as the completion date in these cases. This made the sickness measure more reliable, but mildly underestimated.

Only 83 participants had received medical rehabilitation benefits or disability pension without prior sickness benefits. Almost half of the medical rehabilitation recipients had received sickness benefits. The exclusion of these 83 participants should not have influenced the sickness benefit findings to any great extent. Disability pension was neglected in this study due to the low number of recipients, the young age of our sample, and because it

seemingly consisted of a different group of patients. We therefore found it more reasonable to explore sickness benefits as an outcome. In Norway, 34% of 19–24 year olds are students

National welfare insurance

benefits 2003–2011

Medical benefits N=1194 (29.9%)

Sickness benefits N=1117 (28.0%)

Sickness musculoskeletal

N=320 (8.0%)

Sickness mental health N=185 (4.6%)

Medical rehabilitation

benefits N=90 (2.3%)

Medical rehabilitation mental health N=51 (1.3%)

Disability pension N=33 (0.8%)

Social welfare benefits N=695 (17.4%)

Multisite musculoskeletal pain in adolescence